Incorrect Bitwise Shift of Integer

An integer value is specified to be shifted by a negative amount or an amount greater than or equal to the number of bits contained in the value causing an unexpected or indeterminate result.


Description

Specifying a value to be shifted by a negative amount is undefined in various languages. Various computer architectures implement this action in different ways. The compilers and interpreters when generating code to accomplish a shift generally do not do a check for this issue.

Specifying an over-shift, a shift greater than or equal to the number of bits contained in a value to be shifted, produces a result which varies by architecture and compiler. In some languages, this action is specifically listed as producing an undefined result.

Demonstrations

The following examples help to illustrate the nature of this weakness and describe methods or techniques which can be used to mitigate the risk.

Note that the examples here are by no means exhaustive and any given weakness may have many subtle varieties, each of which may require different detection methods or runtime controls.

Example One

A negative shift amount for an x86 or x86_64 shift instruction will produce the number of bits to be shifted by taking a 2's-complement of the shift amount and effectively masking that amount to the lowest 6 bits for a 64 bit shift instruction.

unsigned int r = 1 << -5;

The example above ends up with a shift amount of -5. The hexadecimal value is FFFFFFFFFFFFFFFD which, when bits above the 6th bit are masked off, the shift amount becomes a binary shift value of 111101 which is 61 decimal. A shift of 61 produces a very different result than -5. The previous example is a very simple version of the following code which is probably more realistic of what happens in a real system.

int choose_bit(int reg_bit, int bit_number_from_elsewhere)
{
  if (NEED_TO_SHIFT)
  {
    reg_bit -= bit_number_from_elsewhere;
  }
  return reg_bit;
}
unsigned int handle_io_register(unsigned int *r)
{

  unsigned int the_bit = 1 << choose_bit(5, 10);
  *r |= the_bit;
  return the_bit;
}
int choose_bit(int reg_bit, int bit_number_from_elsewhere)
{
  if (NEED_TO_SHIFT)
  {
    reg_bit -= bit_number_from_elsewhere;
  }
  return reg_bit;
}

unsigned int handle_io_register(unsigned int *r)
{
  int the_bit_number = choose_bit(5, 10);
  if ((the_bit_number > 0) && (the_bit_number < 63))
  {
    unsigned int the_bit = 1 << the_bit_number;
    *r |= the_bit;
  }
  return the_bit;
}

Note that the good example not only checks for negative shifts and disallows them but also for over-shifts. Not bit operation is done if the shift is out of bounds. Depending on the program, perhaps an error message should be logged.

See Also

Comprehensive CWE Dictionary

This view (slice) covers all the elements in CWE.

Weaknesses without Software Fault Patterns

CWE identifiers in this view are weaknesses that do not have associated Software Fault Patterns (SFPs), as covered by the CWE-888 view. As such, they represent gaps in...

Weaknesses Introduced During Implementation

This view (slice) lists weaknesses that can be introduced during implementation.


Common Weakness Enumeration content on this website is copyright of The MITRE Corporation unless otherwise specified. Use of the Common Weakness Enumeration and the associated references on this website are subject to the Terms of Use as specified by The MITRE Corporation.